PARAMETRIC STUDIES ON URBAN GEOMETRY, AIR FLOW AND … · cooler weather, heat islands can be an...

International Journal on Architectural Science, Volume 6, Number 3, p.114-132, 2005

114

PARAMETRIC STUDIES ON URBAN GEOMETRY, AIR FLOW AND TEMPERATURE R. Priyadarsini and N.H. Wong Department of Building, School of Design and Environment National University of Singapore, Republic of Singapore 117566 (Received 20 January 2005; Accepted 4 January 2006) ABSTRACT In low-latitude cities, urban heat islands contribute to the urban dweller’s summer discomfort and significantly higher air conditioning loads. One of the main reasons for the heat island is poor ventilation. Detailed study of the air flow can help to explore how urban canyon effect can be utilized beneficially by designing the canyons in such a way that more parts of the city can be ventilated, thus optimizing heat extraction. High-rise towers distributed randomly are very common in the Central Business District (CBD) area of Singapore. However, the effect of high-rise towers on the velocity and temperature was not well documented. In addition, air flow patterns as well as the temperature distributions in deep and long canyons have received less attention. Hence, this study is a preliminary investigation of the effect of geometric modifications on the air flow as well as temperature patterns within the urban canyons using numerical simulations. It was found that strategically placing a few blocks of high-rise towers will actually help to enhance the velocity thereby reducing the temperature within the canyon when the wind flow in parallel or perpendicular to the canyon. 1. INTRODUCTION Increasing urbanization and industrialization has caused the urban environment to deteriorate. The urban climate and the environmental efficiency of buildings are influenced by the deficiencies in proper development control [1]. As a consequence of changes in the heat balance, air temperatures in densely built urban areas are higher than the temperatures of the surrounding country. This phenomenon, known as Urban Heat Island (UHI), is a reflection of the totality of microclimatic changes brought about by man-made alterations of the urban surface [2]. In high-latitude cities with cooler weather, heat islands can be an asset in reducing heating loads, but in mid and low-latitude cities, heat islands contribute to the urban dweller’s summer discomfort and significantly higher air conditioning loads. 1.1 Causes of Urban Heat Island The effect of building is considered as one of the main reasons for urban heat island effect. Building masses increase the thermal capacity, which has a direct bearing on the city temperature. They reduce wind speed, and give off heat either directly as form of air-conditioning plants and thermal processes or indirectly as the end result of the degradation of mechanical and electrical activities [3]. Due to the lack of ventilation in urban areas, the city’s ability to flush out pollutants is reduced resulting in heat islands [4]. The heat that is absorbed during the day by the buildings, roads and

other construction in an urban area is re-emitted after sunset, creating high temperature differences between urban and rural areas. The urban heat island phenomenon is due to many factors, the most important of which are summarized as follows [5]: The canyon radiative geometry that

contributes to the decrease in long-wave radiation loss from within the street canyon due to the complex exchange between buildings and the screening of the skyline;

The thermal properties of materials, which increase storage of sensible heat in the fabric of the city;

The anthropogenic heat released from combustion of fuels and animal metabolism;

The urban greenhouse, which contributes to the increase in the incoming long-wave radiation from the polluted and warmer urban atmosphere;

The canyon radiative geometry, which decreases the effective albedo of the system because of the multiple reflection of short-wave radiation between the canyon surfaces;

The reduction of evaporating surfaces in the city, which means that more energy is put into sensible heat and less into latent heat; and

The reduced turbulent transfer of heat from within streets.

International Journal on Architectural Science

115

1.2 The Heat Transfer Phenomenon The heat transfer phenomenon which develops between the buildings and the ambience in the urban environment is extremely complex. This phenomenon can be defined on the basis of three basic parameters [6]: 1. The insolation of the buildings, which is a

direct function of the orientation, the morphology of the building and the shading factor due to opposite buildings and the existing shading devices;

2. The wind flow in the street canyon that depends on the road’s orientation in relation to the prevailing wind direction, the geometric characteristics of the canyon and the temperature conditions on the surfaces of the buildings and the road; and

3. The additional heat emission from local points like the air conditioning systems and the road traffic.

1.3 Surface Albedo The albedo of a surface is defined as its hemispherical and wavelength integrated reflectivity [7]. This definition applies to simple uniform surfaces as well as to heterogeneous and complex ones. Typically, urban albedos are in the range of 0.10 to 0.20 but in some cities these values can be exceeded. Using high albedo materials reduces the amount of solar radiation absorbed through building envelopes and urban structures and keeps their surfaces cooler. A white surface with an albedo of 0.61 was only 5°C warmer than ambient air whereas conventional gravel with an albedo of 0.09 was 30°C warmer than air [7]. One dimensional meteorological simulation by Taha et al. [8] showed that localized afternoon air temperatures on summer days can be lowered by as much as 4°C by changing the surface albedo from 0.25 to 0.4 in a typical mid-latitude warm climate. The differences in energy use between two identical residential buildings one with a moderately light surface with albedo 0.7 and the other with a medium dark colour with albedo 0.3 was simulated for different locations in United States [9]. The results showed how the sensitivity of energy use to colour varies with location, in this case for a decrease in albedo from 0.7 to 0.3. Albedo seems to be effective in determining energy use in hot climates and mitigation of heat island effect and its resulting high cooling loads [8]. 1.4 Urban Canyon The urban canyon is a more useful city unit for investigation in the UHI study. The distribution of the ambient air in a canyon greatly influences the

energy consumption of the buildings [6]. One of the main reasons for heat built up in the heat island is poor ventilation. The wind speed in the canopy layer is seriously decreased compared to the undisturbed wind speed and its direction may be altered [10]. In parallel, the wind and the temperature regime in canyons dramatically affect the potential for natural ventilation of urban buildings and thus the possibility of using passive cooling techniques instead of air conditioning. Detailed study of the air flow can help to explore how urban canyon effect can be utilized beneficially by designing the canyons in such a way that more parts of the city can be ventilated, thus optimizing heat extraction. The urban wind is one that can be dominated and modified by urban design. The main urban design elements which can modify the wind conditions are the overall density of the urban area, size and height of the individual buildings, existence of high-rise buildings and the orientation and width of the streets [11]. Characteristics of canyon geometries, expressed in terms of height-to-width (H/W) and length-to-height (L/H) ratio, are known to produce three principal air flow regimes: ‘isolated roughness’; ‘wake interference’; and ‘skimming flow’ [12]. There have been various studies of the relationship between the canyon geometry, sky view factor, heat island intensity and the surface temperatures [13,14]. The results indicate that the average air temperature of the streets is governed by more complex and regional factors than their surface temperatures. Results from simultaneous measurements performed in three sets of canyon in Athens pointed out that the air temperature measured in the middle of the canyon is not influenced by the orientation of the street either during the day or during the night. This leads to the conclusion that air temperature in the canyon is not greatly influenced by the canyon configuration and is mainly controlled by the air-flow process [1]. High-rise towers distributed randomly are very common in the Central Business District of Singapore (see Fig. 1). However, it was noted that the effect of high-rise towers on the velocity and temperature was not well documented. In addition, air flow patterns as well as the temperature distributions in deep and long canyons have received less attention. Hence, the objective of this study is to investigate the effect of geometric modifications on the air flow as well as temperature patterns within the urban canyons. This study makes use of Computational fluid dynamics (CFD) simulations for the air flow analysis. CFD is a powerful numerical modeling tool for prediction of wind and temperature conditions around buildings. The details of urban canopy can be efficiently taken into account because of the fine grid generation capabilities.


116

CFD provides a cost-effective and accurate alternative to scale model testing. The variations on the simulations can be performed quickly. An appropriate numerical fluid flow model is an ideal tool to assist in the understanding of complex flow problems and to reach at an optimal design scheme from possible alternatives [15]. 2. METHODOLOGY A canyon having a height of 18 m and length of 250 m was selected for the study. This is typical of the canyons found in the Central Business District area of Singapore. No high-rise towers were included in this Model which is named as Model 1 (see Fig. 2). Another model, Model 2 was created by including the high-rise towers to explore the

effect of these towers on the airflow profile. The H/W ratio was varied from 1.5 to 0.4. The wind was allowed to flow parallel as well as perpendicular to the canyon. Wind speed of 4 ms-1, based on the meteorological data was used for this study. The velocities at the middle of the canyon at various Points were considered for analysis. Fig. 2 shows the three Points along the canyon length. Points 1 and 3 are located at the two ends of the canyon where as Point 2 is situated in the middle of the canyon. In Model 2, Point 1 is in between a high-rise tower of 180 m height on one side and low-rise continuous canyon on the other side as shown in Fig. 3. Point 2 is at the middle and does not have any high-rise tower on either side. Point 3 is in between two high-rise towers of height 130 and 150 m.

Fig. 1: View of the street canyons in CBD area

Fig. 2: Plan of Model 1 and measurement locations

Point 1 Point 2 Point 3 Parallel wind flow

Perpendicular wind flow

w


117

Fig. 3: Plan of Model 2 and measurement locations 2.1 Numerical Simulations The numerical simulation was based on the two equation k-ε model using the software CFX 5.6 [16]. The steady state flow has been solved using the basic transport equations for continuity, momentum and energy as shown in Table 1. The inlet boundary condition was expressed in the form of power law wind profile as explained below to simulate the atmospheric boundary layer. Power law:

a

refref ZZ

UU

=

Where, reference height zref = 10 m, coefficient of roughness for city centre a = 0.4.

The air temperature was assumed to be 30°C. The outlet was modeled as free pressure boundary with zero relative static pressure. The surface temperature data of the building façade materials were measured and the values were assigned for the external surface of the buildings as fixed temperature boundary conditions. Here the wall boundary is fixed at a specified temperature Tw. The heat flux into the domain is calculated by qw = hc(Tw-Tnw) where Tnw is the near wall temperature and hc involves the use of turbulent wall functions. The ground surface was assigned a surface temperature value of 45°C to represent the road which is made of asphalt. The solver run was performed until convergence and accurate balance of mass and energy were achieved.

Table 1: Transport equations of the k-ε turbulence model

Continuity 0=

∂∂

i

i

XU

Momentum θβ

ρ ii

i

i

jt

jij

jj g

XU

XUvv

XXXUU −

∂∂

+∂∂

+∂∂

+∂Π∂

−=∂∂ )(1

Thermal energy θ

θσθ

αθ qX

vXX

Ui

t

iii +

∂∂

+

∂∂

=∂∂

Turbulence Kinetic energy εθ

σβ

σ θ

−∂∂

+∂∂

∂∂

+∂

∂+

∂∂

+

∂∂

=∂∂

i

ti

j

i

j

i

i

jt

jk

t

jjj

Xvg

XU

XU

XU

vXkvv

XXkU

Turbulence dissipation rate

i

ti

j

i

j

i

i

jt

j

t

jjj X

vg

kC

kC

XU

XU

XU

vCkX

vv

XXU

∂∂

+−

∂∂

∂∂

+∂

∂+

∂∂

+

∂∂

=∂∂ θ

σβεεεε

σε

ε 03

2

21

Eddy viscosity

)/(2εkCv Dt =

Point 1 Point 2 Point 3

High-rise towers

w


118

3. RESULTS AND DISCUSSIONS 3.1 Parallel Wind Flow 3.1.1 Velocity at Point 1

Fig. 4 shows the velocities at various heights of Point 1 in the case of Model 1 as well as Model 2. For Model 1, the velocities change slightly with the change in H/W ratio. Here, three vertical zones can be distinguished according to the change in behaviour of velocity with the H/W ratio. At lower levels (1.5 m & 5 m), for H/W ratio of 1.5, wind gets deflected by the building corners and thus channeling effect was more prominent. When the H/W ratio was changed to 1, the channeling effect was weaker and the velocity decreases. When H/W ratio was further reduced, the behavior was a kind of unobstructed flow and the velocity increased with the increase in street’s width. At the middle level (10 m), velocity is almost constant. At the upper level (15 m), the channeling effect becomes less prominent. When H/W ratio decreases from 1.5 to 1, the velocity first increases due to an increase in the air mass entering as a result of the street’s width increase. After that the velocity remains almost constant. With further decrease of the H/W ratio to 0.6, a flow separation occurs leaving an area of low velocity at the centre and high velocity at the two sides. The flow becomes unobstructed when the H/W ratio reduced to 0.4, and hence the velocity increases. In the case of Model 2, it can be observed that there is a very high increase in velocity at all the heights compared to Model 1 and the maximum velocities at the lower and middle levels have increased by 114% and 94% respectively. Here again, three vertical zones can be clearly observed. At lower levels (1.5 m & 5 m), the velocity decreases with the decrease in H/W ratio even though the difference is not very significant. This shows that channeling effect is predominant at the base and is decreasing with the increase in street’s width. At the middle level (10 m, 15 m & 30 m), the velocity first increases with the increase in street’s width, reaches a maximum when H/W ratio is 0.8 and then decreases. When the H/W ratio is 1.5, there is a high velocity region at the centre induced by the deflection from the high-rise towers. This condition persists, even more intensely for H/W ratio 1 and 0.8. When the H/W ratio is further reduced to 0.6, this effect is lost at the centre of the canyon and is shifted towards the high-rise tower and hence the velocity is lower. The upper level (60 m ht) also shows a similar behavior as the middle level. However, the velocity increases substantially when

H/W ratio changes from 1.5 to 1, because of the wind deflection from the tower. Thereafter the velocity remains constant when the H/W ratio changes from 0.8 to 0.6 and then decreases when H/W is reduced to 0.4. 3.1.2 Temperature at Point 1

Fig. 5 shows the temperatures at various heights at Point 1. The road with high surface temperature has a direct impact on the air temperature for Model 1. Here, the temperature at the upper level did not change significantly for different H/W ratios. The minimum temperature was always observed for H/W of 1.5. The maximum temperature was up to 31.5°C. Fig. 6 and Fig. 7 show the temperature contours for Model 1 when H/W ratio is 1.5 and 0.6 respectively. Fig. 8 and Fig. 9 show the corresponding velocity vectors for these cases. It can be seen that as the road surface area increases, an upward flow of hot air can be observed. For Model 2, it can be seen that the temperature is substantially lower than Model 1. The maximum temperature was only 30.4°C and this is 1.2°C lower than Model 1. There is no significant difference in temperature for different H/W ratios. However, at 60 m height, the temperature was slightly higher for H/W 1.5. Referring back to the velocity profile in Fig. 4, it can be seen that the velocity is much lower at this location, which shows a negative correlation between velocity and temperature. 3.1.3 Velocity at Point 2

Fig. 10 shows the velocities at Point 2 which is at the centre of the canyon length. For Model 1, it can be seen that lower levels (1.5 m & 5 m ht) are mainly governed by the channeling effect. Here velocity decreases when H/W changes from 1.5 to 0.8 as the channeling effect is reduced after which the velocity increases with the decrease in H/W ratio as a result of the unobstructed flow. At the middle level (10 m ht) channeling effect is less significant. The velocity first increases when H/W reduces from 1.5 to 1 as more air mass flows in. After that the flow is separated into different zones along the length of the canyon, leaving a low velocity area in the centre. Hence the velocity decreases for H/W = 0.8 and 0.6. Then the velocity increases slightly when H/W ratio changes to 0.4 due to unobstructed flow. At the upper level (15 m), the velocity first increases as the increased air mass entering due to the increase in street’s width. The velocities decreases from this point onwards due to the flow separation at the centre.


119

0.51

1.52

2.53

3.54

4.55

5.5

1.5 1 0.8 0.6 0.4

H/W ratio

Velo

city

(m/s

)Model 1_1.5mModel 1_5mModel 1_10mModel 1_15mModel 2_1.5mModel 2_5mModel 2_10mModel 2_15mModel 2_30mModel 2_60m

Fig. 4: Velocities at Point 1 (parallel wind flow)

3030.230.430.630.8

3131.231.431.631.8

1.5 1 0.8 0.6 0.4

H/W ratio

Tem

pera

ture

(°C

) Model 1_1.5mModel 1_5mModel 1_10mModel 1_15mModel 2_1.5mModel 2_5mModel 2_10mModel 2_15mModel 2_30mModel 2_60m

Fig. 5: Temperatures at Point 1 (parallel wind flow)

Fig. 6: Temperature contours for Model 1 at Point 1 for H/W = 1.5 (parallel wind flow)


120

Fig. 7: Temperature contours for Model 1 at Point 1 for H/W = 0.6 (parallel wind flow)

Fig. 8: Velocity vectors for Model 1 at Point 1 for H/W = 1.5 (parallel wind flow)

Fig. 9: Velocity vectors for Model 1 at Point 1 for H/W = 0.6 (parallel wind flow)


121

11.21.41.61.8

22.22.42.62.8

1.5 1 0.8 0.6 0.4

H/W ratio

Velo

city

(m/s

)

Model 1_1.5mModel 1_5mModel 1_10mModel 1_15mModel 2_1.5mModel 2_5mModel 2_10mModel 2_15m


Looking at the velocities for Model 2, it can be seen that the effect of the high-rise tower in the front has an influence on the velocity even at the lower levels (1.5 m & 5 m). The minimum velocity, which was 1.1 ms-1 in the case of Model 1, has been now increased to 2.1 ms-1. When H/W ratio changes from 1.5 to 1, velocity increases as more air mass flows in. With further decrease in H/W ratio, the velocity decreases. Even though there is increased velocity due to deflection from the tower, the high velocity region is shifted away from the centre and towards the tower. At the middle level (10 m), the velocity decreases with decrease in H/W ratio until the H/W ratio is 0.6 after which the velocity increases as a result of unobstructed flow. At the upper level (15 m), when H/W ratio is 1.5, the velocity is high due to the deflection by the tower behind, but this is lost when H/W is reduced to 1. After that the flow is mainly governed by free movement of air and the velocity increases with the decrease in H/W ratio. 3.1.4 Temperature at Point 2

Fig. 11 shows the temperature at different heights of Point 2. For Model 1, the maximum temperature observed was up to 31.6°C and the minimum temperature was up to 30.2°C. Lower temperature was found for high H/W ratio at all the levels. For Model 2, the maximum temperature has been reduced to 30.6°C. Highest temperature was observed at 1.5 m height and the temperature increased with increase in the width of the streets. At the upper level, temperature decreased with increase in H/W ratio.

3.1.5 Velocity at Point 3

Fig. 12 shows the velocities for Model 1 at Point 3 which is near the end of the canyon and farthest

from the inlet. The velocities at this Point is larger compared to the other two Points. For Model 1, all the three zones showed similar profile. However, at 1.5 m height, the profile was a little different. The velocity first increases with the decrease in H/W ratio due to high exit velocity, reaches a maximum when H/W is 0.8. After that a flow separation occurs, leaving low velocity area in the middle of the canyon and hence the velocity decreased considerably for H/W of 0.6. When H/W ratio is changed to 0.4, the velocities at the lower and middle levels increase where as the velocities at the upper level reduce. For Model 2, it is to be noted that this point is in between two high-rise towers. The velocities at the lower and middle levels are 95% higher than Model 1. Here two distinct vertical zones can be observed in terms of the velocity profile. At lower levels (1.5 m & 5 m), the effect of deflection from the two high-rise towers is insignificant. However, channeling effect is enhanced by these towers. When H/W ratio changes from 1.5 to 1, velocity increases as more air mass flows in. The velocity remains constant for H/W ratio of 0.8 and 0.6 and then decreases for H/W 0.4, as a result of the flow separation at the middle. At the middle/upper levels (10 & 15 m / 30 m & 60 m), the influence of deflection from high-rise towers becomes very significant. For H/W ratio of 1.5, the wind gets deflected from the shorter tower thus increasing the velocity. As a result, high velocity can be observed near the longer tower. When the H/W ratio is reduced to 0.8 and 0.6, the velocity at the centre of the canyon is more due to the combined affect of the shorter tower as well as the longer tower. When the H/W ratio is further reduced to 0.4, a flow separation occurs thus resulting in a lower velocity at the centre. Here the spacing between the two


122

tower blocks is one of the important factors governing the flow profile. 3.1.6 Temperature at Point 3

Fig. 13 shows the temperatures at Point 3. For Model 1, when H/W ratio was 0.4, high temperature of 31.8°C was observed at 1.5 m height. Here, minimum temperature was observed for H/W ratio of 1. At other levels, the minimum temperature was obtained for H/W = 0.8. For Model 2, the maximum temperature observed was 30.9°C which is around 1°C lower than Model 1. The temperatures did not show much difference except at 1.5 m height. The temperature at lower zone decreased when H/W ratio decreased from 1.5 to 0.6. However, when the H/W ratio changed to 0.4, the temperature increased due to the influence from the increased road area. 3.1.7 Regression analysis

Multiple linear regression analysis was conducted in order to find out the relation between temperature, velocity and canyon geometry. Temperature (T) at a Point was considered as the dependent variable and the relation with other independent variables like velocity (V), width of the canyon (W) and the vertical distance of the location from the ground (h) was analyzed. Table

2 and Table 3 show the results of multiple linear regression analysis for Model 1 and Model 2 respectively for parallel wind flow. For Model 1, the R2 values are quite high which suggests a good fit. The negative sign of the velocity coefficient shows that the temperature reduces as velocity increases and the positive value of the coefficients for “W” shows that the temperature increased with increase in street’s width. It is expected that as the street’s width increases, there is more airflow which ultimately would reduce the air temperature within the canyons. However, the velocities as well as temperatures at the lower levels were highly influenced by the high temperature (45°C) of the road surface. Parallel flow generates a mean wind along the canyon axis with uplift along the canyon walls as the flow is retarded by friction at the building walls and street surface [17]. For wider canyons, this frictional force leads to an upward air flow at the middle of the road as a result of which the hot air from the road surface rises up. For Model 1, the coefficients for velocity decreased and the coefficients for street’s width increased when moved away from the inlet. Hence Point 3 had the lowest “V” coefficient and highest “W” coefficient. For Model 2, the velocity coefficients are quite low even though the sign is negative. Also the R2 values are not very high. The “W” coefficients are negative for Points 1 and 2, and positive for Point 3.

3030.230.430.630.8

3131.231.431.631.8

1.5 1 0.8 0.6 0.4

H/W ratio

Tem

pera

ture

(°C)

Model 1_1.5m

Model 1_5m

Model 1_10m

Model 1_15m

Model 2_1.5m

Model 2_5m

Model 2_10m

Model 2_15m

Fig. 11: Temperatures at Point 2 (parallel wind flow)


123

11.5

22.5

33.5

44.5

55.5

6

1.5 1 0.8 0.6 0.4

H/W ratio

Velo

city

(m/s

)

Model 1_1.5mModel 1_5mModel 1_10mModel 1_15mModel 2_1.5mModel 2_5mModel 2_10mModel 2_15mModel 2_30mModel 2_60m


3030.230.430.630.8

3131.231.431.631.8

32

1.5 1 0.8 0.6 0.4

H/W ratio

Tem

pera

ture

(°C

)

Model 1_1.5mModel 1_5mModel 1_10mModel 1_15mModel 2_1.5mModel 2_5mModel 2_10mModel 2_15mModel 2_30mModel 2_60m

Fig. 13: Temperature at Point 3 (parallel wind flow)

Table 2: Multiple linear regression analysis for Model 1 for parallel wind flow

(Figure in bracket refers to “t” value)

Independent variables Dependent variable (T) Parallel wind

Point 1 Point 2 Point 3 Constant 32.22 (47.84) 31.88 (58.66) 31.54 (59.17) V (velocity) -1.03 (-1.44) -0.84 (-1.81) -0.67 (-2.22) W (width of streets) 0.009 (2.286) 0.013 (3.44) 0.017 (3.17) H (ht from ground) -0.02 (-0.45) -0.03 (-1.39) -0.02 (-1.36) R-square 0.9 0.84 0.64 F-statistics 75.93 37.77 12.81


124

Table 3: Multiple linear regression analysis for Model 2 for parallel wind flow

Independent variables Dependent variable (T) Parallel wind

Point1 Point 2 Point 3 Constant 30.51 (234.86) 31.62 (93.82) 30.79 (38) V -0.07 (-1.44) -0.57 (-3.7) -0.104 (-0.47) W -0.002 (-1.8) -0.0007 (-0.44) 0.002 (0.82) H -0.005 (-1.57) -0.002 (-0.498) -0.016 (-1.08) R-square 0.7 0.58 0.56 F-statistics 12.22 7.35 6.79

3.1.8 Summary

For parallel wind flow, the velocities at lower levels were enhanced by the channeling effect by the building corners, in the case of canyons with higher H/W ratios. For canyons with lower H/W ratios, unobstructed flow was observed. At higher levels, the channeling effect was found to be not very significant. The velocity first increases with decrease in H/W ratio, and then reduces due to the flow separation at the centre. The channeling effect observed at the lower levels in the absence of the high-rise towers existed for even smaller H/W ratios with the introduction of high rise towers. This means unobstructed flow occurs at larger widths. The velocities at the three Points within the canyon were highly enhanced by the presence of high-rise towers. The velocity was increased by up to 90% and the temperature was reduced by up to 1°C with the introduction of high-rise towers. 3.2 Perpendicular Wind Flow 3.2.1 Velocity at Point 2

Point 2 which is at the centre of the canyon is selected for the analysis. Fig. 14 shows the velocities at Point 2. For Model 1, it can be seen that the velocities are much lower. The maximum velocity observed is less than 0.5 ms-1. Figs. 15 to 19 show the velocity vectors for different H/W ratios. For H/W ratio of 1.5, a weak vortex is formed near the upwind building. For H/W =1 the vortex is more distinct. When H/W is reduced to 0.8, the vortex is at the canyon centre and for H/W = 0.6, vortex is shifted downwards and the characteristics of wake interference flow can also be observed. For H/W of 0.4, the spacing is enough to promote unobstructed flow and the characteristics of isolated roughness flow can be noticed. The results were quite similar to the results of long canyons studied by Hunter et al. [18] about the relationship between canyon geometry and transition from one flow regime to other. For Model 2, the velocities are much higher than Model 1. The maximum velocity that was only 0.45 for Model 1, is increased up to 4.3 ms-1 which

is around 10 times higher. Higher velocities were observed at the lower level compared to higher level. Figs. 20 to 24 show the velocity vectors for different H/W ratios. For H/W of 1.5, skimming flow characteristic was observed. H/W of 1 and 0.8 showed similar patterns. For H/W of 0.6, the flow was still skimming flow. However, the velocities at the middle were much lower. For H/W of 0.4, characteristics of wake interference flow was also observed and the vortex has been shifted towards the upwind building. 3.2.2 Temperature at Point 2

Fig. 25 shows the temperatures at Point 2. For Model 1, the temperatures generally decreased with the decrease in H/W ratio. Figs. 26 to 30 show the temperature contours at Point 2 for H/W = 1.5 and 0.4 respectively in the case of Model 1. For Model 2, the maximum temperature at the lower level is around 1.1°C lower than Model 1 and the maximum temperature at the upper level is around 0.7°C lower than Model 1. Similar to Model 1, the temperatures decreased with decrease in H/W ratio. The velocity at the base of the canyon is strong enough to lessen the temperature near the ground surface. Temperature near the ground was lower and at the middle level was higher. This shows a good negative correlation with the velocity profile as shown in Fig. 14. The temperature contours for H/W = 1.5, 1 and 0.4 are shown in Fig. 28, Fig. 29 and Fig. 30 respectively. 3.2.3 Regression analysis

Table 4 and Table 5 show the multiple linear regression analysis for Model 1 and Model 2 respectively for perpendicular wind flow. Unlike parallel wind, the “W” coefficients were negative for Model 1 which shows that temperature decreased with increase in street’s width. The velocity coefficients were positive for Points 1 and 3. This is because the flow is also influenced by the edge effect at these two points. For Model 2, the velocity coefficient is positive for Point 3 which shows that in between two high-rise towers, the temperature increases with the increase in velocity.


125

When the wind flow is perpendicular to the canyon, there is no external air flowing in between the two high-rise towers. Here, the temperature is higher than the temperature of the rest of the domain. The

buoyancy effect causes an internal air flow at this region and the velocity increases with the increase in temperature.

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

1.5 1 0.8 0.6 0.4

H/W ratio

Velo

city

(m/s

)

Model 1_1.5mModel 1_5mModel 1_10mModel 1_15mModel 2_1.5mModel 2_5mModel 2_10mModel 2_15m

Fig. 14: Velocities at Point 2 (perpendicular wind flow)

Fig. 15: Velocity vectors for Model 1 at Point 2 for H/W = 1.5 (perpendicular wind flow)

Fig. 16: Velocity vectors for Model 1 at Point 2 for H/W = 1 (perpendicular wind flow)


126





127


Fig. 21: Velocity vectors for Model 2 at Point 2 for H/W = 1 (perpendicular wind flow)



128



30.4

30.6

30.8

31

31.2

31.4

31.6

31.8

32

1.5 1 0.8 0.6 0.4

H/W ratio

Tem

pera

ture

(°C

) Model 1_1.5mModel 1_5mModel 1_10mModel 1_15mModel 2_1.5mModel 2_5mModel 2_10mModel 2_15m

Fig. 25: Temperatures at Point 2 (perpendicular wind flow)


129

Fig. 26: Temperature contours for Model 1 at Point 2 for H/W = 1.5 (perpendicular wind flow)

Fig. 27: Temperature contours for Model 1 at Point 2 for H/W = 0.4 (perpendicular wind flow)

Fig. 28: Temperature contour for Model 2 at Point 2 for H/W = 1.5 (perpendicular wind flow)


130

Fig. 29: Temperature contour for Model 2 at Point 2 for H/W = 1 (perpendicular wind flow)

Fig. 30: Temperature contour for Model 2 at Point 2 for H/W = 0.4 (perpendicular wind flow)

Table 4: Multiple linear regression analysis for Model 1 for perpendicular wind flow

Independent variables Dependent variable (T) Perpendicular wind

Point 1 Point 2 Point 3 Constant 31.74 (289.91) 32.14 (272.33) 31.44 (347.13) V 0.22 (1.06) -0.49 (-1.65) 0.25 (1.85) W -0.02 (-4.63) -0.015 (-5.43) -0.02 (-5.56) H -0.04 (-4.29) -0.017 (-2.62) -0.02 (-2.54) R-square 0.82 0.72 0.84 F-statistics 13.13 12.5 29.93

Table 5: Multiple linear regression analysis for Model 2 for perpendicular wind flow

Independent variables Dependent variable (T) Perpendicular wind

Point 1 Point 2 Point 3 Constant 31.1 (220.97) 31.26 (205.44) 30.71 (225.96) V -0.1 (-2.11) -0.09 (-2.6) 0.05 (0.86) W -0.004 (-1.8) -0.009 (-4.38) -0.006 (-3.04) H -0.003 (-0.33) 0.004 (0.81) -0.001 (-0.3) R-square 0.51 0.59 0.39 F-statistics 5.55 7.7 3.43


131

3.2.4 Summary

The velocities as well as temperatures of the lower levels were highly influenced by the high temperature of the road surface. The velocity was increased up to 10 times and the temperature has been reduced by 1.1°C. The temperatures were much higher near the ground. But the introduction of high-rise towers led to strong air flow at the lower level thus decreasing the temperature. The temperature generally decreased with the increase in velocity. 4. CONCLUSION From the series of parametric studies conducted by modifying the urban canyon geometry, it was found that strategically placing a few blocks of high-rise towers will actually help to enhance the velocity within the canyon when the wind flow in parallel or perpendicular to the canyon. In addition, the temperature was lower when high-rise towers were introduced. For parallel flow, the velocity was increased by up to 90% and the temperature was reduced by up to 1°C with the introduction of high-rise towers. For perpendicular flow, the velocity was increased up to 10 times and the temperature was reduced by 1.1°C. However, the transition to wake interference flow occurred at a larger H/W ratio. The velocities as well as temperatures at the lower level were highly influenced by the high temperature of the road surface. However, when the wind flow was perpendicular, the introduction of high-rise towers caused stronger air flow at the lower level thus decreasing the temperature. The channeling effect observed at lower levels of the canyons in the absence of the high-rise towers existed for even smaller H/W ratio with the introduction of high-rise towers. That means unobstructed flow occurred at larger widths. Multiple linear regression analysis showed that the temperature reduces as velocity increases. The positive value of the “W” coefficients in the case of parallel wind shows that the temperature increased with increase in street’s width. The “W” coefficients were negative in the case of perpendicular wind which shows that temperature decreased with increase in street’s width. However, the lengths of the canyon, height of the high rise towers as well as the spacing between the high-rise towers are other important parameters which need further in-depth study. Also, it is to be stressed here that the CFD simulations need to be validated with experimental results.

NOMENCLATURE C1,C2 empirical constant in generation/destruction

term of ε-equation C3 empirical constant in buoyant term of ε-

equation CD empirical constant for eddy viscosity gi gravitational constant in Xi direction (ms-1) k mean turbulent kinetic energy (m2s-2)

iU mean velocity component in Xi direction (ms-1)

jU mean velocity component in Xj direction (ms-1)

Xi, Xj distance in Cartesian coordinate (m) Greek Letters β volumetric expansion coefficient (K-1) ε mean dissipation rate of k (m2s-3) θ temperature rise above ambient α thermal diffusivity v kinematic molecular viscosity (m2s-1) vt eddy viscosity (m2s-1) Π total pressure (Nm-2); Π = ( )( )3/2 kpp + ρ fluid density (kgm-3) σk empirical constant of turbulent Prandtl

number for k σε empirical constant of turbulent Prandtl

number for ε σθ empirical constant of turbulent Prandtl

number for θ REFERENCES 1. M. Santamouris, N. Papanikolaou, I. Livada, I.

Koronakis, C. Georgakis, A. Argiriou, and D.N. Assimakopoulos, “On the impact of urban climate on the energy consumption of buildings”, Solar Energy, Vol. 70, pp. 201-216 (2001).

2. H.E. Landsberg, The urban climate, Academic Press, New York (1981).

3. C.P. Tso and K.Y. Law, “Evaluation of the thermal environment in the building areas of Kuala Lumpur city by a new twin energy balance method”, Energy and Buildings, Vol. 15, pp. 271-277 (1990/91).

4. S. Subbiah, Vasantha Vishwanath and S. Kaveri Devi, “Urban climate in Tamilnadu, India: A statistical analysis of increasing urbanization and changing trends of temperature and rainfall”, Energy and Buildings, Vol. 15, pp. 231-243 (1990/91).

5. T.R. Oke, G.T. Johnson, D.G. Steyn and I.D. Watson, “Simulation of surface urban heat islands under “ideal” conditions at night - Part 2: Diagnosis and causation”, Boundary Layer Meteorology, Vol. 56, pp. 339-358 (1991).


132

6. A.M. Papadopoulos, “The influence of street canyons on the cooling load of buildings and the performance of air conditioning systems”, Energy and Buildings, Vol. 33, pp. 601-607 (2001).

7. H. Taha, “Urban climates and heat islands: albedo, evapotranspiration and anthropogenic heat”, Energy and Buildings, Vol. 25, pp. 99-103 (1997).

8. H. Taha, H. Akbari, A. Rosenfel and J. Huang, “Residential cooling loads and the urban heat island-the effects of albedo”, Building and Environment, Vol. 23, No. 4, pp. 271-283 (1988).

9. C. Conner, “The effect of exterior surface absorptivity and emissivity on energy use in residential buildings”, Unpublished report, Lawrence Berkeley Laboratory (1985).

10. M. Santamouris, Energy and climate in the urban built environment, London: James & James Science Publishers (2001).

11. B. Givoni, Climate considerations in building and urban design, John Wiley & Sons Inc. (1998).

12. T.R. Oke, “Street design and urban canopy layer climate”, Energy and Buildings, Vol. 11, pp. 103-113 (1988).

13. L. Barring, J.O. Mattsson and S. Lindovist, “Canyon geometry, street temperatures and urban heat island in Malmo, Sweden”, Journal of Climatology, Vol. 5, pp. 433-444 (1985).

14. S. Yamshita, K. Sekine, M. Shoda, K. Yamashita and Y. Hara, “On the relationship between Heat island and sky view factor in the cities of Tama river basin, Japan”, Atmospheric Environment, Vol. 20, No. 4, pp. 681-686 (1986).

15. T.T. Chow, Z. Lin and Q.W. Wang, “Applying CFD simulation in analyzing split type air-conditioner performance at buildings”, Architectural Science Review, Vol. 43, No. 3, pp. 133-140 (2000).

16. ANSYS CFX 5.6, User manual.

17. Y. Nakamura and T.R. Oke, “Wind, temperature and stability conditions in an east-west oriented urban canyon”, Atmospheric Environment, Vol. 22, No. 12, pp. 2691-2700 (1988).

18. L.J. Hunter, I.D. Watson and G.T. Johnson, “Modelling airflow regimes in urban canyons”, Energy and Buildings, Vol. 15, pp. 315-324 (1990/91).

PARAMETRIC STUDIES ON URBAN GEOMETRY, AIR FLOW AND … · cooler weather, heat islands can be an...

Documents

Transcript of PARAMETRIC STUDIES ON URBAN GEOMETRY, AIR FLOW AND … · cooler weather, heat islands can be an...